Download Nontraumatic Corneal Perforation

Cornea 19(3): 313–319, 2000.
© 2000 Lippincott Williams & Wilkins, Inc., Philadelphia
Nontraumatic Corneal Perforation
Manapon Lekskul, M.D., Harvey U. Fracht, M.D., Elisabeth J. Cohen, M.D.,
Christopher J. Rapuano, M.D., and Peter R. Laibson, M.D.
Purpose. To study the predisposing conditions, treatments, and
visual outcomes of nontraumatic corneal perforations. Methods. A
retrospective chart review was conducted of all nontraumatic corneal perforations seen between January 1992 and December 1998,
with ⱖ3 months of follow-up, at the Cornea Service Wills Eye
Hospital. Results. A total of 40 nontraumatic corneal perforations
was analyzed. Sixty-two percent of the cases were female. At
presentation, 35 of 40 eyes (87.5%) had best corrected visual acuity of 20/200 or worse. The most common diseases associated with
perforations were keratoconjunctivitis sicca (12 eyes, 30%), bacterial keratitis (6 eyes, 15%), exposure keratopathy (5 eyes,
12.5%), and herpes simplex virus (HSV) keratitis (4 eyes, 10%).
Visual acuity improved ⱖ2 Snellen lines in 3 of 8 eyes (37.5%)
treated with penetrating keratoplasty, 5 of 14 eyes (35.7%) treated
with tissue adhesive, and 1 of 12 eyes (8.3%) given medical treatment. After allowing for the different levels of presenting vision,
treatment modality was not significantly related to final visual
outcome. Conclusion. Keratoconjunctivitis sicca is the most common underlying disease associated with nontraumatic corneal perforation. Corneal perforations were managed successfully using
tissue adhesive, medical therapy, or penetrating keratoplasty.
Treatment depended on the characteristics of the perforation and
on the visual potential of the eye.
Key Words: Corneal perforation—Tissue adhesive—Penetrating keratoplasty.
PATIENTS AND METHODS
The records of patients with the diagnosis of corneal perforation
without a history of eye trauma or corneal surgery who were
treated on the Cornea Service at the Wills Eye Hospital from
February 1992 through January 1999 with at least three months
follow-up, were reviewed retrospectively. If the patients had repeated episodes of corneal perforations in the same eye, only the
first corneal perforation was analyzed. Diagnoses were based on
clinical history, biomicroscopic examination, tear-film studies, as
well as smears and cultures.
The data analyzed included patient age, sex, best corrected visual acuity (BCVA) at presentation, final BCVA, size, location and
shape of corneal perforation, ocular history, and medical treatment.
Treatment and visual outcome also were analyzed. Visual outcomes as they related to varying diagnoses were analyzed by examining the covariance of final visual acuity as the dependent
variable and the baseline visual acuity as the covariate, with treatment as the grouping variable. The Wilcoxon rank sum test for
independent groups was used to analyze the correlation between
age and rheumatoid arthritis (RA).
Histoacryl (B. Braun Melsungen AG D-34209, Melsungen, Germany) was used as a tissue adhesive. It was applied under topical
proparacaine anesthesia after the insertion of a lid speculum. The
surrounding epithelium was scraped, and the area of the perforation was dried with a cellulose sponge. A small amount of tissue
adhesive was applied to the perforation and ulcer, reaching the
surrounding normal basement membrane tissue. A flat, low-watercontent bandage soft contact lens was placed before removal of the
lid speculum.
In this study, we treated corneal perforations with tissue adhesive when the perforations were <1.0 mm in diameter at the level
of Descemet’s membrane and were concave. If active infection
was present, it was treated for 48–72 hours before tissue adhesive
was applied. We managed corneal perforations medically when
they were small (ⱕ1.0 mm in diameter), without iris prolapse, and
could not be treated with tissue adhesive because of ectatic configuration, or when associated with a poor visual prognosis for PK.
Medical treatments used to manage perforations included patching
and/or shielding, topical and systemic hypotensive agents to diminish the flow of aqueous humor, topical and systemic antibiotics
(all received topical and systemic antibiotics), and antifungal
agents (in cases of fungal keratitis).
A PK was used to treat large corneal perforations (>1.0 mm in
diameter) with iris prolapse, with a flat or very shallow anterior
chambers for >48 hours, and with good visual potential that failed
to respond to medical treatment or tissue adhesive. The corneal
button was removed, and a 0.5 mm oversized donor cornea was
Nontraumatic corneal ulceration leading to perforation is the
result of many different noninfectious and infectious destructive
conditions of the cornea. Because corneal perforation has a high
ocular morbidity, prompt recognition and treatment may preserve
useful vision.1 Failure to diagnose and treat a perforation early can
lead to further corneal damage, cataract formation, glaucoma development, endophthalmitis, and loss of the eye. Available treatments for the management of nontraumatic corneal perforation
include medical treatment,2,3 tissue adhesive,4,5 lamellar keratoplasty, and penetrating keratoplasty (PK).1,3,6
Submitted June 25, 1999. Revision received September 1, 1999. Accepted September 3, 1999.
Cornea Service (M.L., H.U.F., E.J.C., C.J.R., P.R.L.), Wills Eye
Hospital, and the Department of Ophthalmology (E.J.C., C.J.R., P.R.L.),
Thomas Jefferson University Medical College, Philadelphia, Pennsylvania,
U.S.A.
Address correspondence and reprint requests to Dr. E.J. Cohen, Cornea
Service, Wills Eye Hospital, 900 Walnut Street, Philadelphia, PA 19107,
U.S.A.; E-mail: [email protected]
All authors have no financial interest in this study.
313
314
M. LEKSKUL ET AL.
TABLE 1. Nontraumatic corneal perforation
Age
(yr)
Sex
1
2
3
4
5
6
7
8
9
10
11
12a
12b
13
14a
14b
15a
15b
16
17
18
19
20
69
65
74
82
82
65
38
39
73
85
45
73
F
F
F
M
F
F
M
F
F
F
F
F
91
74
M
F
71
M
41
54
75
25
55
M
F
M
F
F
21
22
23
24
25
26
49
73
66
55
36
58
M
M
F
F
F
M
CF
1⬘
HM
20/200
20/60
CF
20/400
27
28
29
30
31
32
33
87
41
49
77
79
42
76
F
M
F
M
M
F
F
HM
CF
1⬘
20/60
HM
LP
HM
LP
34
35
36
37
61
76
68
77
M
F
M
F
CF
CF
HM
NLP
No.
Presentation
BCVA
20/200
CF
2⬘
LP
20/200
HM
20/200
20/200
HM
HM
LP
20/40
HM
HM
CF
1⬘
20/200
20/80
20/80
CF
3⬘
20/400
HM
LP
HM
2⬘
4⬘
Diagnosis
History
KCS/RA
Sterile corneal ulcer
KCS
OCP
S. aureus corneal ulcer
KCS/Sjögren’s syndrome
HSV keratitis
HSV keratitis
Pseudomonas + S. aureus corneal ulcer
S. aureus corneal ulcer
Rosacea
KCS/Sjögren’s syndrome
Corneal ulcer 3 wk; hypothyroid
Corneal ulcer 4 wk
HZO 4 mo
Epith. defect 1 mo
Recurrent HSV keratitis 12 yr
HSV keratitis 7 yr
Corneal ulcer 3 wk, HZO 9 yr + recurrent trichiasis
HZO 1 yr
Seasonal allergy
RA
Entropion
KCS
KCS/Sjögren’s syndrome
RA, blepharitis
Rosacea
Fungal (filament, Alterania species) corneal ulcer
Exposure keratopathy
Corneal ulcer (neg. culture)
Exposure keratopathy/neuroparalysis
keratopathy
Exposure + neuroparalysis keratopathy
KCS/Sjögren’s syndrome
Exposure keratopathy
Chronic staph, hypersensitivity
Corneal ulcer (neg. culture)
Coag. neg. Staphylococcus + S. viridans
corneal ulcer
OCP
HSV keratitis
KCS/Sjögren’s syndrome
Exposure kertopathy
HZO
Fungal (candida) corneal ulcer
Fungal (filament Chrysonilia species)
corneal ulcer
HSV keratitis
KCS
S. epidermidis + S. viridans corneal ulcer
S. viridans corneal ulcer
sutured using a 10-0 nylon interrupted suture technique, after peripheral iridectomies were performed. Occasionally, a lamellar
keratoplasty was performed using full-thickness corneal donor tissue but leaving some host tissue peripheral to the perforation in the
HZO 6 yr, neurotrophic keratopathy
Ptosis repair, NVG, hypothyroid
Down syndrome, hypothyroid
CVA, facial palsy
Radiation for maxillary sinus adenocarcinoma
Final BCVA
20/25
20/40
LP
20/100
20/200
20/30
20/200
HM
HM
LP
20/30
CF
HM
HM
20/400
CF 3⬘
20/80
20/70
20/60
LP
NPL
NLP
Thyroidectomy
Hypothyroid
NIDDM, ulcerative colitis
Rosacea, eczema, atopic dis.
CF 1⬘
CF 6⬘
20/400
20/400
NLP
20/80
Recurrent HSV keratitis 20 yr
RA
Lid surgery or skin cancer
Temporal arteritis 8 mo
Corneal ulcer 3 wk, eczema
HZO, viral hepatitis carrier
NLP
20/200
20/60
NLP
LP
LP
LP
Recurrent keratouveitis, cirrhosis, lymphadenopathy
HZO 7 yr
Lid surgery 1 yr/blepharitis
20/30
20/200
HM
NLP
recipient bed. We avoided conjunctival flaps in patients with corneal perforations. Generally, the size of the perforation and visual
potential, but not the location (peripheral vs. central), determined
the management.
FIG. 1. Age and sex distribution of
patients.
Cornea, Vol. 19, No. 3, 2000
NONTRAUMATIC CORNEAL PERFORATION
TABLE 2. Initial and final best corrected visual acuity
RESULTS
Visual acuity
Initial (eyes)
Final (eyes)
20/20–20/40
20/50–20/100
20/200–20/400
Count fingers
Hand motion
Light perception
No light perception
Not recorded
1
4
8
8
12
5
1
1
5
6
7
4
5
6
6
1
TABLE 3. Causes of nontraumatic corneal perforation
Diagnosis
Male
Female
Total
KCS
Bacterial keratitis
Exposure keratopathy
HSV keratopathy
Fungal keratitis
OCP
Rosacea
HXV keratopathy
Entropion
Chronic staphylococcal hypersensitivity
Corneal ulcer (unknown cause)
2
2
3
3
0
1
1
1
1
0
0
7
4
2
1
3
1
1
0
0
1
3
9
6
5
4
3
2
2
1
1
1
3
KCS, keratoconjunctivitis sicca; HSV, herpes simples virus; OCP,
ocular cicatricial pemphigoid; HZV, herpes zoster virus.
TABLE 4. Organisms isolated on culture
Organism
Number
Bacteria
Staphylococcus aureus
Streptococcus viridans
Staph. coagulase negative
Pseudomonas
Fungus
Filamentous fungi
Candida
315
3
3
2
1
2
1
A total of 40 eyes of 37 patients were treated for nontraumatic
corneal perforations over a seven-year period (Table 1). There
were 14 men (38%) and 23 women (62%). The ages ranged from
25 to 91 years, with a mean of 63.4 years (Fig. 1). The follow-up
period ranged from 3 to 84 months (mean, 24.5 months).
At presentation, 35 eyes (87.5%) had a BCVA of ⱕ20/200
(Table 2). One patient had Down’s syndrome, and visual acuity
could not be tested accurately. The location of the perforation was
central in 22 eyes and peripheral in 18 eyes.
The most common cause of perforation was keratoconjunctivitis
sicca (KCS) in 12 eyes of nine patients. Other causes included bacterial keratitis (six eyes), exposure keratopathy (five eyes), HSV keratitis (four eyes), fungal keratitis (three eyes), ocular pemphigoid (two
eyes), and rosacea (two eyes) (Table 3). Cultures or scrapings (Table
4) confirmed the presence of bacterial or fungal infections.
The underlying medical diseases included secondary Sjögren’s
syndrome related to RA (five patients), thyroid disease (two patients with exposure keratopathy and three on thyroid medication
only), and rosacea (three patients), as well as primary Sjögren’s
Syndrome (one patient). Nine perforations were secondary to infection (six bacterial and three fungal). Bacterial isolates included
Staphylococcus, Streptococcus, and Pseudomonas. Fungal isolates
included both yeasts and filamentous organisms (Table 4). There
was no statistically significant correlation between age and RA
(Wilcoxon rank sum test for independent groups, p ⳱ 0.45). Furthermore, we found no significant relationship between the average visual improvement in perforations associated with RA versus
those related to another diagnosis after adjustment for different
treatment modalities.
Tissue adhesive was applied in 15 eyes. Of the 15 eyes treated
with tissue adhesive, 8 had to be reglued for recurrent leaks or glue
dislodgment within several days and 1 eventually needed PK for
refractory leaking. Thirteen eyes were treated by medical treatment. The wound sealing time for medically treated leaks ranged
TABLE 5. Final best corrected visual gain or loss
Diagnosis
KCS
OCP
Bacterial
Fungal
HSV
HZV
Rosacea
Entropion
Exposure
Staph. hypersen.
Corneal ulcer (unknown cause)
Treatment
No. (eye)
Medical
Tissue adhesive
Tissue adhesive → PK
Medical
Medical
PK
Medical
PK
Tissue adhesive
PK
LK
Medical
Tissue adhesive
PK
Tissue adhesive
PK
Eviscerationa
Tissue adhesive
Medicalb
PK
Enucleationa
2
9
1
2
5
1
2
1
2
2
1
1
1
1
1
2
2
1
1
1
1
Gain ⱖ2
Gain 1
0
Loss 1
Loss ⱖ2
4
1
1
1
2
1
1
1
1
1
1
1
NLP
1
3
2
1
2
1
1
1
1
1
1
1
1
1
2
1
1
1
a
Blind painful eye.
Down syndrome.
KCS, keratoconjunctivitis sicca; HSV, herpes simplex virus; OCP, ocular cicatricial pemphigoid; HZV, herpes zoster virus; PK, penetrating
keratoplasty; LK, lamellar keratoplasty.
b
Cornea, Vol. 19, No. 3, 2000
316
M. LEKSKUL ET AL.
TABLE 6. Treatment and final best corrected visual outcome
Treatment
Number
Gain ⱖ2
Gain 1
0
Loss 1
Loss ⱖ2
NLP
Medical
Tissue adhesive
Tissue adhesive → PK
PK/LK
12
14
1
(8/1)
1
5
0
3
4
1
0
1
4
4
1
2/1
2
1
0
2
0
2
0
0
1
1
0
0
PK, penetrating keratoplasty; LK, lamellar keratoplasty; NLP, no light perception.
from 2 to 30 days, with a mean of 8.8 days. Nine eyes developed
repeated perforations. These subsequent perforations were excluded from further evaluation. PK was used when the perforation
could not be glued and the visual potential was good. PK was
performed in eight eyes. Of these eight, four eyes (50%) improved
in BCVA and four eyes achieved a clear graft. Lamellar keratoplasty was used in one eye. In cases of RA, medical management
included systemic immunosuppression in some patients. Two
FIG. 2. BCVA at presentation vs. final BCVA. (A) Combined graph. (B) Medical treatment.
Cornea, Vol. 19, No. 3, 2000
NONTRAUMATIC CORNEAL PERFORATION
blind painful eyes were eviscerated, and one blind painful eye was
enucleated after successful treatment of the perforations (Table 5).
The initial and final BCVA demonstrated a statistically significant
direct relationship for all therapies combined.
Visual acuity improved ⱖ2 Snellen lines in 3 of 8 eyes (37.5%)
treated with PK, in 5 of 14 eyes (35.7%) of those treated with
tissue adhesive, and in 1 of 12 eyes (8.3%) of those treated medically (Table 6). The final BCVA after the treatment of 15 (37.5%)
of nontraumatic corneal perforations was improved (Fig. 2).
317
DISCUSSION
In this study, the etiologies of nontraumatic corneal perforations
included sterile corneal ulcers (23 patients, 57.5%) that are most
commonly associated with a) KCS and RA, b) exposure keratopathy, c) neurotrophic keratopathy (HSV in 4 patients and herpes
zoster virus in 1 patient), as well as bacterial and fungal corneal
ulcers (9 patients, 22.5%). KCS is the most common cause of
nontraumatic corneal perforation in this study (12 of 40 eyes,
FIG. 2. BCVA at presentation vs. final BCVA. (C) PK. (D) Tissue adhesive.
Cornea, Vol. 19, No. 3, 2000
318
M. LEKSKUL ET AL.
30%). Women were more frequently affected by this problem than
men (seven of nine patients, 77.8%). We found RA to be the most
common underlying condition associated with KCS-related corneal perforation (five of nine patients [55.6%], four women and
one man). The ages ranged from 49 to 73 years, with a mean of
65.4 years. RA is the most common collagen vascular disorder that
involves the ocular surface.7 Women are primarily affected. Of RA
patients, 24–31% have Sjögren’s syndrome.8
When a patient presents to an ophthalmologist with a corneal
perforation associated with KCS, advanced RA is usually evident.
Keratolysis is the mechanism of corneal perforation in these patients.9 In most cases, severe aqueous tear deficiency is found. The
superficial layers of the cornea begin to ulcerate and a descemetocele may develop and be followed by perforation.10 Dry eyes
should be treated aggressively in RA patients. Frequent lubrication, punctal occlusion, and tarsorrhaphy may be useful in promoting reepithelialization. Immunosuppressive agents may be useful
in alleviating the immunologic process responsible for the corneal
melting.11
This study demonstrated a statistically significant direct correlation between presenting and final BCVA for the aggregate of all
treatment types (R2 ⳱ 0.38, p ⳱ 0.0001, Fig. 2). That is, the better
the presenting vision, the better the final vision after therapy. We
found no statistically significant correlation between treatment
modality and final visual acuity after adjusting for varying presenting visual acuity (Figs. 2A–C). However, biases existed in this
study. The underlying disease and visual potential did influence
the treatment modality. Consequently, the varying presenting visual acuity, associated with different diagnoses, confounded the
outcome of therapy.
The objective in treating a patient with a corneal perforation is
to restore the integrity of the globe and useful vision. This goal
may require a sequence of procedures. The goals of initial intervention are to close the perforation and restore the integrity of the
globe. This should be done as rapidly as possible to minimize
peripheral anterior synechiae and risk of cataract formation and
intraocular infection.
Tissue adhesive works well when the corneal perforation is
ⱕ1.0 mm in diameter at the level of Descemet’s membrane, away
from limbus, and is concave in shape with a crater for the tissue
adhesive. In many cases, adhesive is sufficient treatment that promotes healing and scarring while obviating the need for further
surgery. Tissue adhesive is left in place for several months until it
dislodges spontaneously. Adhesive has been shown to be bacteriostatic to gram-positive (but not to gram-negative) organisms,
especially during polymerization.12 Tissue adhesive has also been
shown to slow stromal-melting,13 possibly by preventing invasion
by tear-borne inflammatory cells. Adhesive supports the stroma
tectonically through vascularization and fibroplasia.14 Tissue adhesive is not an option for large corneal perforations (>2 mm in
diameter) or in anterior bulging corneal perforations.
In our experience, the application of tissue adhesive was often
sufficient to promote adequate healing without surgery. Many surgeons recommend keratoplasty following tissue adhesive applications.15 In this study, tissue adhesive was applied in 15 eyes; 14
eyes (93.3%) achieved tectonic stability and 6 (42.8%) improved in
final BCVA. Only 1 of 15 eyes required a PK because the perforation was not sealed by tissue adhesive.
Medical treatments can be effective in the management of small
perforations (ⱕ1 mm in diameter) without iris prolapse. However,
Cornea, Vol. 19, No. 3, 2000
in these cases, the time to resolution of the leak may be prolonged.
Medical therapy was used when application of tissue adhesive was
not feasible because of an ectatic perforation and/or active infection. In pinpoint or small perforations, patching may be effective if
stromal swelling seals the perforation and there is no infection. We
used topical or systemic hypotensive agents to diminish the flow of
aqueous humor and to promote closure of the tissue defect.16
Using medical treatment, we were able to achieve tectonic stability in 13 eyes (32.5%). Final BCVA was improved in 5 of 12
eyes (41.6%). Visual acuity was not measured in one case because
the patient had Down’s syndrome. If there was no contraindication
to medical treatment, we used this treatment for infectious corneal
perforation first because the outcome of surgical treatment is suboptimal when the infection is uncontrolled. Perforations are also
frequently ectatic in this setting.
Larger central perforations (>1 mm in diameter) should be
treated with PK as initial therapy if the visual potential is good. We
believe that patients with microbial keratitis should be treated aggressively with topical fortified antibiotics for a period of 24–48
hours before PK. Corneal transplantation in a perforated cornea is
technically difficult, and the possibility of injuring other anterior
segment structures is always present. Although the anatomic success in our keratoplasty series was high, only four of eight (50%)
achieved a clear graft. The success of PK in the treatment of
corneal perforation depends on the timing of surgery and the cause
of the perforation. Ideally, PK is deferred until the eye is quiet.
However, this is frequently not possible when the perforation cannot be glued or does not respond to medical therapy. If the visual
acuity after tissue adhesive was poor because of scarring, a PK
could be performed at a later date when the inflammation had
subsided. However, in our experience, this was usually not necessary. Corneal transplantation was not recommended in patients
with perforations successfully managed medically or with tissue
adhesive because of the poor prognosis associated with severe
ocular surface disease and, in some cases, preexisting posterior
segment pathology.
Finally, we can decrease the incidence of nontraumatic corneal
perforation by early detection and aggressive treatment of ocular
surface conditions including KCS, HSV, and herpes zoster virus
neurotrophic keratopathy, exposure keratopathy, and rosacea. In
addition, these underlying diseases must be treated after perforation to decrease the likelihood of recurrent problems and to improve the outcome of PK.
REFERENCES
1. Partnoy SL, Insler MS, Kaufman HE. Surgical management of corneal
ulceration and perforation. Surv Ophthalmol 1989;34:47–58.
2. Arentsen JJ, Laibson PR, Cohen EJ. Management of corneal descemetoceles and perforations. Ophthalmic Surg 1985;16:29–33.
3. Lin TLD, Webster RG, Abbot RL. Repair of corneal lacerations and
perforations. Int Ophthalmol Clin 1988;28:63–75.
4. Hirst LW, Smiddy WE, Stark WJ. Corneal perforations: changing
methods of treatment, 1960–1980. Ophthalmology 1982;89:630–5.
5. Moschos M, Droutsas D, Boussalis P, Tsioulias G. Clinical experience
with cyanoacrylate tissue adhesive. Doc Ophthalmol 1997;93:237–45.
6. Nobe JR, Moura BT, Robin JB, Smith RE. Results of penetrating
keratoplasty for the treatment of corneal perforation. Arch Ophthalmol
1990;108:939–41.
7. Koffler D. The immunology of rheumatoid disease. Ciba Clin Symp
1979;31:1–10.
NONTRAUMATIC CORNEAL PERFORATION
8. Robin JB, Dugel R, Robin SB. Immunologic disorders of the cornea
and conjunctiva. In: Kaufman HE, Barron BA, McDonald MB. The
cornea, 2nd ed. Boston: Butterworth-Heinemann, 1998:571–3.
9. Jayson MIV, Easty DL. Ulceration of the cornea in rheumatoid arthritis. Ann Rheum Dis 1977;36:428–35.
10. Kervick GN, Pflugfelder SC, Haimovici R, Brown H, Tozman E, Yee
R. Paracentral rheumatoid corneal ulceration: clinical features and cyclosporine therapy. Ophthalmology 1992;99:80–8.
11. Palay DA, Stulting RD, Waring III GO, Wilson LA. Penetrating keratoplasty in patients with rheumatoid arthritis. Ophthalmology 1992;99:
80–8.
319
12. Maguen E, Nesburn AB, Macy JI. Combined use of sodium hyaluronate and tissue adhesive in penetrating keratoplasty of corneal perforations. Ophthalmic Surg 1984;15:55–7.
13. Fogle JA, Kenyon RK, Foster CS. Tissue adhesive arrests stromal
melting in human cornea. Am J Ophthalmol 1980;89:795–802.
14. Kenyon RK. Corneal perforations: discussion. Ophthalmology 1982;
89:634–5.
15. Saini JS, Sharma A, Grewal SPS. Chronic corneal perforations. Ophthalmic Surg 1992;23:399–402.
16. Mannis MJ, Ruben J, Wedemeyer L. Corneal fistulas and their management. Am J Ophthalmol 1988;105:626–31.
Cornea, Vol. 19, No. 3, 2000